pygit2 blob历史记录

Question

我正在尝试使用pygit2在git bare存储库中做等同于git log filename的工作。文档只解释了如何执行这样的git log：

from pygit2 import GIT_SORT_TIME
for commit in repo.walk(oid, GIT_SORT_TIME):
    print(commit.hex)

你有什么想法吗？

谢谢

编辑：

我现在有这样的东西，或多或少准确地说：

from pygit2 import GIT_SORT_TIME, Repository


repo = Repository('/path/to/repo')

def iter_commits(name):
    last_commit = None
    last_oid = None

    # loops through all the commits
    for commit in repo.walk(repo.head.oid, GIT_SORT_TIME):

        # checks if the file exists
        if name in commit.tree:
            # has it changed since last commit?
            # let's compare it's sha with the previous found sha
            oid = commit.tree[name].oid
            has_changed = (oid != last_oid and last_oid)

            if has_changed:
                yield last_commit

            last_oid = oid
        else:
            last_oid = None

        last_commit = commit

    if last_oid:
        yield last_commit


for commit in iter_commits("AUTHORS"):
    print(commit.message, commit.author.name, commit.commit_time)

Answer 1

我建议您只使用git的命令行界面，它可以提供格式化良好的输出，使用Python进行解析非常容易。例如，要获取给定文件的作者姓名、日志消息和提交散列：

import subprocess
subprocess.check_output(['git','log','--pretty="%H,%cn%n----%B----"','some_git_file.py'])

有关可以传递给--pretty的格式说明符的完整列表，请查看git日志：https://www.kernel.org/pub/software/scm/git/docs/git-log.html的文档

Answer 2

另一种解决方案是，懒惰地从给定的提交中生成文件的修订版。由于它是递归的，如果历史记录太大，它可能会崩溃。

def revisions(commit, file, last=None):
    try:
        entry = commit.tree[file]
    except KeyError:
        return
    if entry != last:
        yield entry
        last = entry
    for parent in commit.parents:
        for rev in revisions(parent, file, last):
            yield rev